128

Earnings Management through Goodwill Impairment:

CEO and CFO tenure impact Olga Vladimidrovna Viśnevskaá MSc. LL.B.1

Executive summary

This paper examines the relationship between the extent of goodwill impairment and the properties of CEOs and CFOs of a selection of FTSE Eurotop 100 Index companies. Prior research indicated that it is likely that CEOs tend to take earnings baths early in their tenure, as the losses can then still easily be blamed on their predecessors, as well as creating a lower benchmark for measuring their own future financial performance. Also, the nature of a specific turnover process and the prior employment of the incoming CEO (hired from within or outside the company) have been considered as an explanatory variable by some studies. The outcomes of this study indicate that the tenure and prior employment of the CEO are significantly associated with a company’s financial reporting behavior in relation to the magnitude of goodwill impairment. However, contrary to expectations, goodwill impairment charges are likely to increase as the tenure of a CEO increases. CEOs promoted from inside the same company are likely to impair goodwill by larger amounts, compared to CEOs hired from outside the company. A significant association between the CFO tenure and prior employment variables and the magnitude of impairment charges was not established in this study. For the full text of this master thesis refer to the following webpage: http://hdl.handle.net/2105/5550. 1. Introduction The objective of this study is to asses the extent of goodwill impairment by European companies for the period 2006-2007, and to investigate the relationship between the extent of goodwill impairment and the properties of executives in charge at the time. The commonly used opportunistic perspective of the Positive Accounting Theory predicts that when self-interested actors are confronted with opportunities to use discretion with

1 This thesis was supervised by Drs. C.D. Knoops. Olga Viśnevskaá has graduated from the Erasmus School of

Economics and is currently in completing her masters thesis on the topic of financial law at the Rotterdam

School of Law.

129

regard to financial accounting and reporting to their own advantage, they will do so. This practice of ‘earnings management’ could be aimed at either increasing the reported income, or decreasing it through income smoothing and taking of earnings baths. The International Financial Reporting Standards are often criticized for allowing room for discretion especially due to the prescription of use of fair values. More specifically, the accounting treatment of goodwill through the use of impairment tests is often criticized. My own examination of the financial reporting standards revealed that indeed, in my opinion, there was room for managerial discretion with respect to goodwill (re)valuation and possible losses arising from it. Accordingly, I expect executives to use goodwill impairment charges to manage earnings to achieve personal goals. As I wonder whether personal goals could be related to the phase of employment of an executive, I formulate the following research question:

Are tenure and prior employment of the CEO and the CFO associated with a company’s financial reporting behavior in relation to the magnitude of goodwill impairment?

Considering the prior research mentioned further in this master thesis, this study mainly builds and expands on the work conducted by Masters-Stout e.a. 2007. My study adds value

to the existing body of research for the following reasons:

• Firstly, contrary to most studies mentioned in this master thesis, as well as that by Masters-Stout, this study is conducted using data of European companies that are subject to IFRS and not SFAS. The outcomes can thus be considered more relevant in the European context;

• Secondly, as far as my knowledge goes, no other study has been conducted on the relationship between the CFO tenure and prior employment and a company’s financial reporting behavior regarding the magnitude of goodwill impairment;

• Finally, as far as I know, no other study has combined and offset both CEO and CFO properties in relation to goodwill impairment, in one research design.

The remainder of this article is organized as follows. First, I will describe the theoretical background to my study and review the outcomes of prior research on the subject. Continuingly, I will introduce the hypotheses that were tested and follow with a brief description of the sample used in my study. I will then describe the research design and the corresponding model and continue with the elaboration of the results. Before I conclude this article, I will reflect on the outcomes of my study and give suggestions for further research. 2. Theoretical background and prior literature

2.1 Earnings management and financial reporting incentives

The practice of managers trying to influence the financial reporting numbers and the way they appear in the financial statements is often known by the term ‘earnings

130

management’. Several incentives to manage earnings can be identified, for example (Palepu e.a., 2007):

1. accounting-based debt covenants: requirement of certain debt-contracts and meeting targets arising from them, can induce managers to distort accounting figures to gain more favorable results;

2. management compensation: (bonus-)compensation which are often connected to reported profits and wanting to secure their position for longer period of time, is another motivation to favorably influence the reported income;

3. corporate control contests: managers can use accounting numbers to gain approval of company’s shareholders in their attempt to become/remain a manager.

There are several ways for managers to influence financial reporting, one of which is asset distortion. When managers desire to increase reported earnings they tend to overstate assets, as this is accompanied by either an increase in income or a reduction of costs in the income statement. On the other hand, managers can also desire to deflate earnings by understating assets. Managers can ‘smooth income’ by overstating expenses during a period of exceptional performance by the company. Also managers can ‘take a bath’ in income by overstating expenses during a period of exceptionally bad performance to create an appearance of a turnaround in the following years (Palepu e.a., 2007). Thus, managers are not necessarily interested in presenting accounting figures only ‘for the better’ as earnings baths occur as well. 2.2 Discretionary financial reporting and the case of goodwill The first step in the examination of prior research concentrated on the studies of the relationship between opportunistic behaviour and goodwill impairment testing. Prior research showed evidence of opportunistic behavior on the account of managers with regard to impairment testing of goodwill as the prescribed accounting treatment. The findings were, however, not uniform. Some researchers (Anantharaman 2007, Henning and Shaw 2004) found little support for the criticism of goodwill impairment testing, which was introduced as the new accounting treatment of goodwill in SFAS 142 and IAS 36. To the contrary, other studies have shown evidence of the misuse of managerial discretion to some degree under the new accounting standards (Beatty and Weber 2005, Lapointe 2005, Zang 2008, Ramanna and Watts 2007, Carlin e.a. 2007). 2.3 Management tenure: the role of the chief executive officer Like any process, the period of tenure of the chief executive officer (CEO) can be divided into different phases. The agency theory predicts that managers are guided by self-interest. Presumably, different phases of tenure will correspond with different goals and motivations. And so, the second step of the literature examination was aimed at examining the relationship between the executive management’s phase of employment and its influence on financial reporting. Moore (1973) conducted one of the first studies on the subject of the influence of management changes in the field of accounting. He found that in the year of a top management change, income reducing discretionary accounting decisions, such as write-

131

downs, write-offs and taking of provisions, occurred significantly more than in years with no management change. He interpreted the overall results to be an indication of the newly appointed management taking an earnings bath. More so, because the majority of companies with indication of income-reducing discretionary accounting decision did report an increase in income in the first reporting year after the change. Accordingly, Moore hypothesized that the incentives of the incoming management for taking the income reducing discretionary decisions are two-fold. First, the blame would be placed on their predecessors and the historical benchmark for their own future performance is reduced. Second, the losses taken in the year of the change would not have to be reported in the future, thus increasing the future reported income and the appearance of their performance. Later, DeAngelo (1987) found that when a ‘dissident’ (an outside manager) was hired, he would report an ‘immediate earnings bath’, so to be able to report an earnings turn-around in the following years. Pourciau (1993) investigated the behavior of incoming managers in cases of what she called a non-routine (involuntary) executive change. She found that for these instances the incoming executives managed accruals in the year of the change to reduce income, and did the opposite in the following year. Additionally, in the year of the change, larger write-off were taken. Francis e.a. (1996) conducted a broad study of possible causes of discretionary write-offs. Among others, she found that write-offs occur more frequently if preceded by a management change, and are then also larger in size. Several studies, which did not directly investigate the relationship between executive tenure and goodwill impairment, did produce outcomes on this subject as well. Like Beatty and Weber (2005), who hypothesized that the difference between actual and predicted goodwill write-offs could be explained by the departure of the CEO who made the original acquisition decision. Further, the study by Lapointe (2005) also found that higher transitional goodwill impairment losses correlated with companies having experienced recent management change. Additionally Ramanna and Watts (2007) found that goodwill-write offs are negatively associated with CEO tenure. Finally, Zang (2008) found that recent management change was an explanatory variable for earnings management through transitional goodwill impairment losses2, as he believed that higher goodwill impairment losses were taken during the transitional period to increase the likelihood of higher earnings in the future. Bengtsson e.a. (2007) investigated the occurrence of earnings management in Sweden, surrounding management turnovers through both accruals, as well as write-offs. Earnings were reduced in the first year of the turnover and increased in the following year. This supports the findings in the previously mentioned studies. Furthermore, Bengtsson attempted to distinguish an association between earnings management and an executive turnover in question, being routine versus non-routine. However, he found no conclusive evidence in support of this distinction.

2 Loss incurred by companies upon the adoption of the new SFAS. 142 standard.

132

Masters-Stout e.a. (2007) performed a subsequent study, which related goodwill impairments under SFAS 142 to CEO tenure. For the companies that did impair, she found that newly appointed CEOs reported higher impairments than senior CEOs.’ Her additional findings indicated that CEOs hired internally from within the present employees of the company impaired relatively smaller amounts. She hypothesized that these CEOs were more ‘personally invested’ in previously taken strategic acquisition decisions and thus lacked what she called a ‘fresh perspective’. These outcomes were however insignificant. 2.4 Management tenure: the role of the chief financial officer From the previous step of my prior research analysis it became clear that scholars hypothesize that CEOs have certain incentives to manipulate financial reporting, have the power to do so, and use their power to act on their incentives. In the continuing step I attempt to consider the role of another senior manager, which could be presumed to influence the financial reporting of a company: the chief financial officer (CFO). The role of the modern CFO is no longer limited to mere ‘financial record keeping’. Now, the CFO ‘is one of the top decision makers – often leading member of the top management along with the chief executive officer and the chief operating officer.’ (Copeland, 2001). A CFO today, is involved in decision-making on many levels and about many significant issues throughout the entire company. Intuitively, it can be supposed that some incentives that drive CEOs, might similarly drive CFOs. If so, the agency theory predicts that the CFO will also try to exert influence to satisfy his self-interests. Surprisingly, very few studies have been conducted on the influence of the CFO in the field of accounting research. Could it be more commonsense to consider the CFO to have more influence on financial reporting? More than the CEO? Building on that intuition, Jiang and Petroni (2008) were interested in finding the answer to the question of ‘who has the most influence on earnings management’, the CEO or the CFO. They executed three previously conducted studies, which already established an association between CEOs’ equity incentives and earnings management, and reexamined them by also testing the association between the CFOs’ equity incentives and earnings management. The general outcomes indicated that the amount of discretionary accruals was more closely associated with the CFO rather then with the CEO incentives and that the role of a CFO is indeed influential with regard to a company’s financial reporting behavior. Greiger and North (2006) also suspected that the CFO ‘has a substantial amount of control over a company’s reported financial status’, as they studied the effect of a CFO change on reported accruals. They found that after an appointment of a new CFO, earnings are significantly reduced through the management of accruals. Furthermore, these findings did not seem to be influenced or mitigated by the appointment of a new CEO. Finally, they also found that the hiring of a CFO from a different source than the company’s direct audit company, produced more significant outcomes.

133

3. Hypothesis development

Based on these outcomes of prior research and the predictions of the positive accounting theory I have arrived at the following hypotheses to be tested in my study:

H1: Shorter CEO tenure corresponds with higher goodwill impairment charges. H2: Companies with CEOs, who have been employed by the same company two years or less, will take relatively higher goodwill impairment losses.

Additionally, I have asked myself whether the logic that has been applied to the relationship between properties of a CEO of a company and its financial reporting behavior, could also be applied to the properties of the CFO. Combined with the outcomes of studies regarding the CEO properties mentioned above I arrive at the following additional hypotheses:

H3: Shorter CFO tenure corresponds with higher goodwill impairment charges. H4: Companies with CFOs, who have been employed by the same company two years or less, will take relatively higher goodwill impairment losses.

4. Sample and data collection My study examined the financial data of 58 major European companies listed in the FTSE Eurotop 100 Index during the period 2006-2007, resulting in 116 observations. In 37% of the cases, goodwill impairments were observed. Largest average absolute and relative (measured against revenues) impairment losses were observed in the telecommunications industry. The average CEO tenure for companies within the sample was 5.9 years; the average CFO tenure was 4.5 years. Additionally, 67% of the CEOs in the sample, prior to their appointment, were employed by the same company for less than three years, which classified them as ‘internal hires’ for the purpose of my study. 54% of the CFOs were classified as internal hires. When examining the subsample of companies that that have taken a decision to impair goodwill, the frequency of the impairment decision decreased as observed tenures of CEOs increased. A similar pattern was observed between the frequency of impairment decisions and CFO tenure. The financial data was hand collected using the information provided in the annual reports. Information regarding the tenure and prior employment of the executives was hand-collected for each executive from additional sources like company websites and newspaper articles, as a general database for such information of European companies does not exist. 5. Research design and model To test my hypotheses I used a multivariate regression model, which I will describe in this section.

134

The dependent variable of this model (IMPit) is the reported goodwill impairment charge. First, the effects of the independent variables were measured against the reported nominal impairment amount. Secondly, I believed that there would be added value to measuring the dependent variable relative to the effect this has within the entire income statement of a the specific company, as this puts the impairment charge amount into perspective. This is why, the alternative dependent variable metric is the impairment loss divided by the revenue. To test my hypothesis I added several independent variables of interest. To test H1 and H3, I departed from Master-Stout e.a. (2007), who used dummy variables to distinguish between the new and old executives , as I do not find their arguments to be substantial enough to justify the separation into those specific categories. Using a continuous metric would, as I believed, provide for a test of a more nuanced relationship between the dependent variable and this independent variable of tenure. Furthermore, other studies that have used executive tenure as an independent variable (e.g. Ramanna, Watts, 2007), have also used a continuous metric. Consequently, to test the effect of executive tenure, I use the tenure duration measured in years (CEO_TENUREit and CFO_TENUREit). In line with the hypotheses I have formulated, I expected there to be a negative association between these variables and the dependent variable. To test H2 and H4 I added dummy-variables into the model (CEO_INTERNALit, CFO_INTERNALit). These dummy variables made it possible to test for the difference in the impact of an executive prior employment on the (relative) size of the impairment charge. The dummy variable was coded 0 and is considered to be an external-hire, if the executive in question was employed by company i for less than three years before appointment as an executive officer. In the other case the variable was coded 1. I based this distinction on Master-Stout e.a. (2007), as it seems reasonable to consider an executive who has been with a company for less than three years not to be entrenched. In line with the hypotheses I have formulated, I expected there to be a negative association between these variables and the dependent variable. Additionally, I included several control variables associated with the economic condition of the companies. I used EBITDA (EBITDAit) to control for the size of the economic activity of a company. I saw the EBITDA amount as the measure of the ability of a company to absorb impairment charges. I did not hypothesize a coefficient sign, as, on one hand, I can imagine that higher EBITDA can be seen by the management as buffer that can absorbed ‘unwanted’ expenses and smooth income. On the other hand, lower (than expected) EBITDA could also induce an earnings bath strategy. Further, I included the after tax net income (INCOMEit), which is also used by Masters-Stout e.a. (2007), as a measure of economic performance (profitability) of the companies in the sample. This variable is included in the model to account for the overall profitability of a company. Although net income already includes any impairment losses, I presumed that if a company is confronted with a negative or extremely low or high net income before the publication of final financial results, the management might feel tempted to adjust the reported impairment charge. Consequently, similar pattern that I described for the EBITDA can also be applied to the relationship between the net income and the goodwill impairment

135

charge: high profits could induce income smoothing and losses can be seen as an opportunity to take (further) earnings baths. Thus, no coefficient sign was hypothesized either. The size of a company (SIZEit) measured as the natural logarithm of company’s total assets to normalize the impact of the part of the sample on the larger side of the spectrum, was included in the model as well. I predicted a positive relationship between the size of a company and the amount of the impairment charge, which is also supported by Van de Poel e.a. (2008) outcomes. It seems to me that, as larger companies are often the product of several prior mergers, this would result in more recognized goodwill that in turn can be a subject to impairment. Finally, I included a company’s leverage, measured as total liabilities divided by the total assets, as control variable. I expected highly leveraged companies to be subjected to more attention and scrutiny by their creditors (who are professional investors), and these companies to operate under bigger restrictions of loan contracts. This should in turn reduce the amount of earnings management in general. And when earnings management would take place, it would probably be directed upwards to increase reported earnings. For this reasons I expected a negative coefficient sign for this variable. This variable can be considered a proxy for room for discretionary earnings management by a company as was used by Zang (2008). Hence, my empirical model looks as follows:

I tested this model using the data from the entire sample, as well as performing separate tests on the data from the subsample of companies that impaired goodwill during the observed period. Furthermore, I used two different definitions of the independent variable IMP (measured in absolute and in relative values), and also of the independent variable CEO_TENURE (measured in years and as a natural logarithm). This has resulted in different outcomes. Secondly, I executed three types of regression with regard to the executive properties data. First, I applied the abovementioned model, removing the CFO variables and accounting only for the CEO properties, to measure the ‘pure’ association between IMP and CEO tenure and prior employment. After that, I did the same now removing the CEO tenure and prior employment properties, to measure the ‘CFO effect’. Thirdly, I tested my main model, mentioned above, which controlled the associations between impairment charges and one executive type tenure and prior employment, for that of the other. 6. Results

6.1 CEO properties

The outcomes regarding the independent variables of interest, the CEO properties, are consistent in all cases. However, they are entirely not as expected. CEO tenure is

136

positively associated with impairment charges. Furthermore, compared to CEOs hired from outside the company, CEOs hired from inside the company are associated with higher

impairment charges. These associations are significant at α<.05 for both the entire sample and the subsample, when tenure is expressed as a logarithm. When tenure is expressed in years, measuring impairment relative to revenues, gives slightly better significance results. Overall, these finding indicate that H1 and H2 of my research design are false. These findings are summarized in table 1. Table 3 Regression outcomes CEO properties

6.2 CFO properties

The coefficients of the independent variables of interest were insignificant at α<.05 for all tests performed on the data from the entire sample. Within the subsample, the association between prior employment of the CFO and the magnitude of impairment charges is significant. For this population, compared to CFOs hired from outside the company, CFOs hired from inside the company are associated with relatively higher impairment charges, contrary to expectations. As such, the validity of H3 and H4 was not established for the entire sample, and validity of H3 was not established for the subsample either. H4 was proven to be significantly false for the data of the subsample. These findings are summarized in table 2.

137

Table 4 Regression outcomes CFO properties

6.3 Combined model

With respect to independent variables of interest, several associations become clear. The association between the CEO tenure and the magnitude of impairment charges is

significant at α<.05 in all cases, except for the test of the subsample using the absolute values of IMP and using the tenure variable expressed in years. When tests are performed

on the data of the entire sample CEO prior employment is significant at α<.05, except when IMP is measured in relative values and the CEO tenure is measured in years. Overall, this means that when the decision not to impair is taken into account and the model controls for the association of CFO properties, the association between the CEO tenure and the magnitude of impairment charges is positive. Furthermore, compared to CEOs hired from outside the company, CEOs hired from inside the company are associated with higher impairment charges. Thus, H1 and H2 were proven to be false for the entire sample. In the subsample, the association between the CEO tenure and the magnitude of

impairment charges is significant at α<.05, except for when IMP is measured in absolute values and tenures are measured in years. The association with CEO prior employment is not significant within the subsample. As such, H1 has proven to be false within the subsample, and validity of H2 has not been established. The association between CFO tenure and the magnitude of impairment charges when

controlled for the associations of CEO properties, remains insignificant at α<.05 in all cases. The validity of H3 thus is not established. Furthermore, the association between the magnitude of impairment charges and CFO prior employment is insignificant when analyzing the data of the entire sample. However, within the subsample, CFO prior tenure

association is significant at α<.05, when impairment charge is measured in absolute values, regardless of the definition of tenure. The absolute size of goodwill impairment charges is positively associated with a CFO being promoted from inside, when the decisions not to

138

take goodwill impairment losses is disregarded. H4 is thus proven to be false within the subsample. H4 validity has not been established for the entire sample. These findings are summarized in table 3. Table 5 Regression outcomes controlled for both types of executives

7. Analysis of the outcomes

The first question I ask myself based on the outcomes of my study, is: why would impairment charges increase during the course of employment of a CEO? More specifically, what incentives could there be for managers to want to increase impairment charges late, instead of early, in their tenure? An alternative explanation to wanting to take earnings baths early in their tenure, could be the job security argument. Contrary to the arguments behind my hypotheses, it is conceivable that CEOs might want to show good results (immediately) after their appointment and would want to avoid ‘unnecessary’ losses, to justify their appointment and secure their position. This desire might even induce ‘upward earnings management’. CEOs, who have acquired ‘relational goodwill’ for their positive performance throughout the course of their tenure, could also believe that this would be sufficient to mitigate any harm to their reputation from losses taken in later stages of their tenure, and thus would take these losses easier than ‘younger’ CEOs. Furthermore, the influence of CEO employment contracts and compensation schemes, which is not taken into account in this study, could have alternative explanatory power for the established association between the magnitude of impairment charges and CEO tenure. These contracts and payment schemes are usually constructed (in line with the agency theory) in such way as to align management incentives with company’s/shareholders best interests. Presuming that earnings management is motivated by management self-interest and that employment contracts are constructed effectively, this should lead to a reduction

139

of earnings management. Similarly, Beatty and Weber (2005) found that managers who are subject to more binding ‘contracts that include effects of accounting changes…will prefer to delay expense recognition’. Including executive compensation (plans) into a research model could be a consideration for future research. The second question that I ask myself is: why do the outcomes of my study indicate that internally promoted CEOs impair relatively more than CEOs hired from outside the company? First of all, it is important to note that the majority (69%) of CEOs whose data were included in this study, where classified as internal hires. This fact, by itself, might have a distortive effect on the outcomes of the study. Additionally, perhaps when it comes to employment history, the magnitude of the goodwill impairment charges is not best explained by the duration of CEO prior employment by the same company before his appointment. Reconsidering my prior hypothesis, I believe that the assumption that a manager becomes personally involved in prior acquisition due to his mere presence in the same company, might be too general. Instead, entrenchment could better be defined by the fact whether the manager in question was actually involved in the decision process that preceded a specific acquisition. Some support for this idea was also found in Beatty and Weber’s study (2005), which linked the likelihood of a SFAS 142 impairment to the likelihood of a CEO making ‘the original acquisition’. Again, future studies can inquire to the feasibility of researching prior employment from this angle. An alternative explanation could also be that the prior employment of an executive could represent experience and knowledge. An executive, who has been with the same company for a longer period of time, is likely to have specific inside knowledge that would allow him to make a better judgment about the value of goodwill, and in turn might make it ‘easier’ to take an impairment charge compared to a counterpart who lacks similar experience and knowledge. This could result in the observed relationship between prior employment and the size of impairment charges. Thirdly, I ask myself what other issues there might be that might have influenced or limit the outcomes of my study. First, there is the fact that contrary to most of prior research that has been done on the subject of goodwill impairment and/or the influence of CEO properties (which often have been performed in the United States), my study is aimed at European companies. For example, this could account for the existence of cultural differences between my and the prior research. Perhaps, the European context and tradition with regard to expectations and the regulation of executive behavior, contribute to a less competitive executive environment, which results in less opportunistic behavior. Furthermore, there could be other issues with regard to the chosen sample. For instance there is a noticeable presence of former state owned companies in my sample. As such the oil & gas and the telecommunication sectors combined, account for almost a quarter of the companies in the sample. One can speculate whether these companies are subject to a specific kind of (government-like) corporate culture, which most probably does not exist in the American context. These companies might also still be subject to governmental

140

influence and specific regulation, which would reduce the amount of discretion available to managers. In addition, the descriptive statistics reveal that that the telecommunications industry accounts for the largest goodwill impairment losses observed within my sample. It is a commonly known fact that during the observed period companies in that particular industry underwent several economic adversities, such as the devaluation of purchased UMTS frequencies. The heterogeneity problem surfaces with regard to this issue. It is possible that there were real economic causes to the impairment losses taken in the telecommunications industry. As these are the most significant impairment losses within my sample, this could undermine the validity of the detected relationship between impairment charges and CEO properties. To control for this problem, I ran an additional regressions, which included both CEO and CFO properties, on my data, while removing the entire telecommunication sector. The association between both absolute and relative impairment charges, and CEO tenure remained significant at α < .05 (also, when the decision not to take impairment charges was disregarded). However, the fit of the model measured in R2, decreased significantly to levels lower than 0.10. The CEO prior employment variables and both CFO variables were insignificant. The heterogeneity problem could be addressed through the expansion of the sample and the amount of observations. These are of course the obvious limitations of my study, as it does cover only two years worth of financial data of a limited number of companies. This is an inherent consequence of the nature of the intensive data hand-collection process with regard to the information about CEO tenure and prior employment in the European context. Furthermore, the sample could be expanded with regard to the amount of companies in it. A matter for future research is to consider an entire different sample of companies, or consider including financial institutions in the sample. Finally, based on my own analysis of reporting standards with regard to goodwill impairment and predictions formed in prior research about the susceptibility of goodwill impairment testing to managerial discretion, I formed expectations about goodwill impairment charges to be a likely item to be used for earnings management. This assumption can also be a subject for critical review. It could be a subject to future research to consider to what degree it is really likely that goodwill impairment test is used as a ‘tool’ to manage earnings. Perhaps, other ‘gaps’ in financial reporting standards are used relatively more often to manage earnings (on a larger scale), and these ‘earnings management tools’ could also be tested for association with executive tenure and employment information. 8. Summary and conclusions The outcomes of this study indicate that the tenure and prior employment of the CEO are significantly associated with a company’s financial reporting behavior in relation to the magnitude of goodwill impairment. Contrary to my expectations I have found that CEO

tenure is positively associated with the magnitude of goodwill impairment charges. These

141

results mean that the hypothesis that CEOs tend to take earnings baths in the early stages of their tenure, so losses can more easily be blamed on their predecessors, is false for the data in my sample. Surprisingly, this outcome contradicts the association described in the prior research such as Moore (1973), DeAngelo (1987), Pourciau (1993), Francis e.a. (1996), Lapointe (2005), Ramanna and Watts (2007), Zang (2008), Bengtsson e.a. (2007), Masters-Stout e.a. (2007). Additionally, compared to CEOs hired from outside the company, internally hired CEOs correspond with lager goodwill impairment charges. This falsifies the second type of hypotheses of my thesis that compared to their counterparts, internally hired executives would impair goodwill by smaller amounts, as they are more ‘personally invested’ in previously taken strategic acquisition decisions, and thus would lack a ‘fresh perspective’. This outcome is less surprising as the results of prior research on this topic were inconclusive (Pourciau 1993, Bengtsson e.a. 2007, Masters-Stout e.a. 2007). Contrary to my expectation, I have not established a significant association between the CFO tenure and prior employment variables and the magnitude of impairment charges. At best, I can say that, if the decision not to take impairment charges is disregarded and only the data of the remaining subsamples is tested, CFO prior employment is significantly associated with the magnitude of impairment charges. Within the subsample compared to CFOs hired from outside the company, internally hired CFOs are associated with larger impairment amounts.

References

Alfredson, K., Leo, K., Picker, R., Pacter, P., Radford, J. and V. Wise. 2007. Applying International Financial Reporting Standards. Milton: Wiley and Sons. Anantharaman, D. 2007. Has SFAS 142 deminished the reliability of goodwill write-offs? A PhD-dissertation, draft dated 8 April 2007. Columbia Business School. Beatty, A. and J. Weber. 2005 Accounting discretion in fair values estimates: an examination of SFAS 142 Goodwill Impairments. Journal of Accounting Research 44(2): 257-288. Beckman, H. 2008. Verantwoording van goodwill bij overnames: purchased goodwill dan wel full goodwill. Maandblad voor accountancy en bedrijfseconomie (January/February): 15-24. Bengtsson, K., Bergström, C. and M. Nilsson. 2007. Earnings management and CEO turnover. Working paper. Stockholm School of Economics. Carlin, T.M., Finch, N., Ford, G. Goodwill Impairment – An Assessment of Disclosure Quality and Compliance Levels by Large Listed Australian Firms. Working paper. Macquarie Graduate School of Management

142

Copeland, T.E. 2001. The expanding role of the CFO. Weekly Toyo Keizai September 1. DeAngelo, L.E. 1987. Managerial competition, information costs and corporate governance: the use of accounting performance measures in proxy contests. Journal of Accounting and Economics 10: 3-36. Deegan, C. 2003. Financial Accounting Theory. Roseville: McGraw-Hill. Field, A. 2009. Discovering statistics using SPSS. London: SAGE Publishing. Francis, J.J., Hanna, J.D. and L. Vincent. 1996. Causes and effects of discretionary asset write-offs. Journal of accounting Research 34: 117-134. Gabarro, J.J. 1987. The dynamics of taking charge. Boston: Harward Business School Press. Greiger, M.A. and D.S. North. 2006. Does hiring a new CFO change things? An investigation of changes in discretionary accruals. The Accounting Review 81(4): 781-809. Henning, S.L. and W.H. Shaw. 2004. The amount and timing of goodwill write-offs and revaluations: evidence from U.S. and U.K. firms. Review of Quantitative Finance and Accounting 23: 99-121. Hogendoorn, M.N. and V.C. Hartman. 2007. Fusies en acquisities in de jaarrekening. Fusies en acquisities: fundamentele aspecten van fusies en acquisities. Ed. Frederikslust, van, R.A.I. Amsterdam: LexisNexis Group Europe&Africa. Jensen, M.C and W.H. Meckling. 1976. Theory of the firm: managerial behaviour, agency costs and ownership structure’. Journal of Financial Economics 3(4): 305-360. Jiang, J. and K.R. Petroni. 2008. CFOs and CEOs: who has the most influence on earnings management?. Working paper, draft 15 April 2008. Michigan State University. Johnson, L.T. and K.R. Petrone. 1998. Is goodwill an asset? Accounting Horizons 9: 293-303. Klaassen, J. and J. Van Helleman. 2004. Goodwill. Externe verslaggeving: in theorie en praktijk. Ed. Hogendoorn, M.N. The Hague: Reed Business Information. Knoops, C.D. and C.E. de Bruijn. 2000. Goodwill bij acquisities: percepties van het management van ondernemingen. Maandblad voor accountancy en bedrijfseconomie (August): 308-323. Lapointe, P. 2005. Transitional goodwill impairment losses: economic determinants, reporting incentives and constraints. A PhD- dissertation, draft dated January 2005. Concordia University Montreal.

143

Masters-Stout, B., Costigan, M.L. and L.M. Lovata. 2007. Goodwill impairments and chief executive officer tenure’. Critical Perspectives on Accounting, doi:10/j..cpa.2007.04.002 Moore, M.L. 1973. Management changes and discretionary accounting decisions. Journal of Accounting Research 11(1): 100-107. Palepu, K. G., Healy, P.M., Bernard, V.L. and E. Peel. 2007. Business analysis and valuation: IFRS edition. London: Thomson Learning. Pourciau, S. 1993. Earnings management and nonroutine executive changes. Journal of Accounting and Economics 16: 317-336. Ramanna, K. and R.L. Watts. 2007. Evidence on the effects of unverifiable fair-value accounting. Working paper, draft 31 August 2007. MIT Sloan School of Management. Schipper, K. 1989. Commentary on Earnings Managament. Accounting Horizons 3: 91-102. Van de Poel, K., Maijoor, S. and A. Vanstraelen. 2008. Implementation of IFRS within Europe: the case of goodwill. Working paper, draft dated August 2008. Van Oers, B. 2008. Impairment, is it recoverable? A research on impairment of fixed assets and the quality of disclosed information on impairment in the income statement. Master thesis. Rotterdam: Erasmus University. Zang, Y. 2008. Discretionary behavior with respect to the adoption of SFAS 142 and the behavior of security prices. Review of Accounting and Finance 7(1): 38-68. International Financial Reporting Standard 3: Business Combinations. International Accounting Standards Board. Amended up to 17 January 2008. International Accounting Standard 36: Impairment of Assets. International Accounting Standards Board. Amended up to 17 January 2008.

144

Appendix: companies in the sample

Company Name:

Ceo Name: Ceo appointment year:

Ceo prior employment:

Cfo Name: Cfo appointment year:

Cfo prior employment:

Arcelor Mittal Lakshmi N. Mittal

1989 inside Aditya Mittal

2004 inside

A.P. Moller - Maersk

Jess Søderberg

1994 inside Søren Thorup Sørensen

2006 outside

Air Liquide Benoît Potier 2001 inside John Glen 2001 outside

Anglo American Tony Trahar 2000 outside René Médori

2005 outside

Astra Zeneca David Brennan 2005 inside Jon Symonds

2005 outside

BASF Jürgen Hambrecht

2003 inside Kurt Bock 2003 inside

Bayer AG Werner Wenning

2002 inside Klaus Kühn 2002 inside

BG Group Frank Chapman

2000 inside Ashley Almanza

2002 inside

BHP Billiton Chip Goodyear

2003 inside Alex Vanselow

2006 inside

BMW Group Norbert Reithofer

2006 inside Stefan Krause

2002 inside

BP John Browne 1995 inside Byron Grote

2002 inside

British American Tobacco

Paul Adams 2004 inside Paul Rayner

2002 inside

Carrefour José Luis Durán

1990 inside Eric Reiss 2005 inside

Danone Franck Riboud 2006 inside Antoine Giscard d’Estaing

2005 outside

Deutsche Post Klaus Zumwinkel

1990 outside Edgar Ernst 1995 inside

Deutsche Telekom

René Obermann

2006 inside Karl-Gerhard Eick

2004 inside

Diageo Paul Walsh 2000 inside Nick Rose 1999 inside

EADS Noël Forgeard 2005 inside Hans Peter Ring

2002 outside

Electricite de France (EDF)

Pierre Gadonneix

2004 outside Daniel Camus

2002 outside

Endesa Rafael Miranda Robredo

1997 outside Jose Luis Palomo Alvarez

1991 inside

Enel Fulvio Conti 2005 inside Claudio Machetti

2005 inside

145

Ericsson Carl-Henric Svanberg

2003 outside Karl-Henrik Sundstroem

2003 inside

Company Name:

Ceo Name: Ceo appointment year:

Ceo prior employment:

Cfo Name: Cfo appointment year:

Cfo prior employment:

ENI Paolo Scaroni 2005 outside Marco Mangiagalli

2001 inside

France Telecom Didier Lombard

2005 inside Philippe Jeunet

2000 inside

GlaxoSmithKline Jean-Pierre Garnier

2000 inside Julian Heslop

2005 inside

Iberdrola José Ignacio Sanchez Galán

2001 outside José Sáinz Armada

2002 outside

InBev Carlos Brito 2004 inside Felipe Dutra

2005 inside

Inditex Pablo Isla Álvarez de Tejera

2005 outside Antonio Rubio Merino

2006 inside

L'Oreal Jean-Paul Agon

2005 outside Christian Mulliez

2003 outside

LVMH Bernard Arnault

1989 inside Jean-Jacques Guiony

2004 outside

National Grid Roger Urwin 2001 inside Steve Lucas

2002 inside

Nestlé Peter Brabeck-Letmathe

1997 inside Paul Polman

2006 outside

Nokia Olli-Pekka Kallasvuo

2006 inside Richard A. Simonson

2004 inside

Novartis Daniel Vasella 1999 inside Raymund Breu

1996 inside

Reckitt Benckinser

Bart Becht 1995 inside Colin Day 2000 outside

Repsol Antonio Brufau Niubó

1997 outside Fernando Ramírez Mazarredo

2006 outside

Rio Tinto Tom Albanese 2006 inside Guy Elliott 2002 inside

Roche Group Ranz Humer 1998 inside Erich Hunziker

2001 outside

Royal Dutch Shell

Jeroen van der Veer

1997 inside Peter Voser 2005 outside

Royal KPN A.J. Scheepbouwer

2001 inside M.H.M. Smits

2004 outside

Royal Phillips Electronics

Gerard Kleisterlee

2001 inside Pierre-Jean Sivignon

2005 outside

RWE Harry Roels 2003 outside Klaus Sturany

1999 outside

SAB-Miller Graham Mackay

1999 inside Malcolm Wyman

2001 inside

Saint-Gobain Pierre-André 2005 inside Benoît 2005 inside

146

de Chalendar Bazin

Company Name:

Ceo Name: Ceo appointment year:

Ceo prior employment:

Cfo Name: Cfo appointment year:

Cfo prior employment:

Sanofi-Aventis Jean-Francois Dehecq

1999 inside Jean-Claude Leroy

2004 inside

Suez Gérard Mestrallet

2001 inside Philippe Jeunet

2000 inside

Telecom Italia Riccardo Ruggiero

2002 outside Enrico Parazzini

2001 outside

Telefonica Cesar Alierta 2000 inside Santiago Fernández Valbuen

2002 inside

Telia Sonera Anders Igel 2002 outside Kim Ignatius

2000 outside

Tesco Terry Leahy 1997 inside Andrew Higginson

1997 outside

Total Thierry Desmarest

1995 inside Robert Castaigne

1994 inside

Unilever Patrick Cescau

2005 inside Rudy Markham

2000 inside

Vivendi Jean-Bernard Lévy

2005 inside Jacques Espinasse

2002 outside

Vodafone Arun Sarin 2003 inside Andy Halford

2005 inside

Volkswagen Bernd Pischetsrieder

2002 inside Hans Dieter Pötsch

2003 outside

Xstrata Mick Davis 2001 outside Trevor Reid 2002 outside